Journal of the Neurological Sciences 382 (2017) 116121

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

journal homepage: www.elsevier.com/locate/jns

Free fatty acid as a determinant of ischemic lesion volume in nonarterial- MARK

origin embolic stroke
Jong-Won Chunga,b, Woo-Keun Seoa, Gyeong-Moon Kima, Chin-Sang Chunga, Kwang Ho Leea,
Oh. Young Banga,b,
a
Department of Neurology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
b
Translational and Stem Cell Research Laboratory on Stroke, Samsung Medical Center, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: Goals: This study aimed to determine whether the plasma levels of free fatty acid (FFA) are associated with
Fatty acids ischemic lesion characteristics in nonarterial-origin embolic stroke.
Nonesteried Materials and methods: We prospectively recruited 254 patients with acute cerebral infarction caused by car-
Arrhythmias dioembolic stroke (CES, n = 121) or with embolic stroke of undetermined source (ESUS, n = 133). Plasma
Cardiac
levels of FFA were measured during the acute stage (median of 2 days after stroke onset). Acute ischemic lesions
Embolism
on diusion-weighted imaging were measured in terms of size, composition, and pattern. Transthoracic echo-
Paradoxical
Cerebral infarction cardiography parameters were evaluated in all patients.
Findings: Plasma levels of FAA were not dierent in CES and ESUS patients (mEq/L, 0.78 0.52 vs.
0.67 0.61, P = 0.120). Echocardiography parameters, including left atrium volume index and E/e', were
higher, and the ischemic lesion volume was larger in patients with CES than in those with ESUS. The ischemic
lesion volume and the proportion of patients with mixed (small and large) and large cortical lesions increased
with FFA quartile in both CES and ESUS groups. In a multivariable analysis, FFA level (coecient, 5.249;
standard error, 3.447; P = 0.001), atrial brillation (coecient, 7.673; standard error, 1.855; P < 0.001), and
fasting glucose (coecient, 0.104; standard error, 0.023; P < 0.001) were associated with ischemic lesion
volume in nonarterial-origin embolic stroke.
Conclusion: Elevated plasma FFA levels are associated with larger ischemic lesion volumes and a higher pre-
valence of large cortical infarcts in patients with nonarterial-origin embolic stroke regardless of the presence of a
high-risk cardioembolic source.

1. Introduction
Ischemic stroke is a heterogeneous disease and occurs due to a
multitude of underlying pathologic processes. Cerebral embolic strokes
comprise the majority of ischemic stroke cases. Besides arterial origin
emboli, cardioembolic stroke (CES) is responsible for about 20% [1] of
all cases and embolic stroke of undetermined source (ESUS) stroke
accounts for 2340% of all strokes [24]. Stroke classication schemes
for ischemic stroke etiological subtypes have focused on the detection
of high-risk embolic sources, such as prolonged rhythm monitoring for
paroxysmal atrial brillation. In patients with ESUS and low/uncertain-
risk cardioaortic sources of cerebral embolism, there have been eorts
to measure the likelihood of the stroke being related to these sources
[5]. These eorts include the use of the 10-point Risk of Paradoxical
Embolism (RoPE) score and thorough history taking for pathogenic
patent foramen ovale [6,7], and plaque morphology, size, or mobility
determination for complex aortic arch atheromas [8].
However, there is a possibility that common mechanisms underlie
the progression and development of the two major types of embolic
stroke of nonarterial origin, CES and ESUS. Several decades ago, free
fatty acids (FFAs) were linked to thrombus formation by promoting
platelet aggregation and activating clotting factors in the early stages of
blood coagulation [9,10]. Elevated FFA levels have also been associated
with insulin resistance [11], myocardial dysfunction [12], and ab-
normal cardiac rhythm [13]. Most recently, an association between FFA
and CES has been suggested and recognized as an outcome predictor in
atrial brillation-associated stroke [14,15].
We hypothesized that plasma FFA levels are associated with acute
ischemic lesion characteristics, i.e., ischemic lesion volumes, patterns,
and compositions in patients with nonarterial-origin embolic stroke.

Corresponding author at: Department of Neurology, Samsung Medical Center, Sungkyunkwan University, 81, Irwon-Ro, Gangnam-gu, Seoul, Republic of Korea.
E-mail address: ohyoung.bang@samsung.com (O.Y. Bang).

http://dx.doi.org/10.1016/j.jns.2017.09.040
Received 31 August 2017; Received in revised form 6 September 2017; Accepted 27 September 2017
Available online 30 September 2017
0022-510X/ 2017 Published by Elsevier B.V.
J.-W. Chung et al.
The association was compared between patients with cardioembolism
and those with ESUS. Furthermore, the two types of embolic stroke
were evaluated in terms of their clinical, laboratory, and echocardio-
graphic characteristics.
2. Methods
2.1. Study population
Between May 2014 and December 2015, patients who were ad-
mitted to a university medical center within 7 days of symptom onset
for the treatment of rst-ever nonarterial-origin embolic stroke were
recruited to participate in this study. Patients with high primary cardiac
sources of cerebral embolism, as assessed using the Stop Stroke Study
Trial of Org 10172 in Acute Stroke Treatment (SSS-TOAST) [5], were
placed in the CES group. Patients with (1) ischemic lesion(s) that is not
lacunar, (2) absence of extracranial or intracranial atherosclerosis
causing 50% luminal stenosis or occlusion of arteries supplying the
area of ischemia, (3) no major-risk cardioembolic sources of embolism
by history, electrocardiography (ECG), echocardiography, and 24 h
of cardiac rhythm monitoring, and (4) no other specic causes of is-
chemic stroke (e.g., arteritis, dissection, migraine/vasospasm, or drug
abuse) were placed into the ESUS group [16]. Local Institutional Re-
view Boards approved this study. All patients or their next of kin pro-
vided informed consent to participate in the study.
2.2. Workups
Clinical information, including age, sex, and vascular risk factors
(hypertension, diabetes, hypercholesterolemia, history of coronary ar-
tery disease, and smoking) were acquired systematically. All patients
underwent standardized extensive embolic source diagnostic tests that
included routine blood tests, and cardiac workups, including repetitive
ECGs, transthoracic or transesophageal echocardiography (TEE), echo
bubble tests or transcranial Doppler (TCD) right-to-left shunt tests, and
cardiac telemetry (72 h or more). The TCD right-to-left shunt test is
based on the intracranial detection of intravenously injected micro-
emboli. The size and functional relevance of a right-to-left shunt can be
assessed using TCD with higher sensitivity and more predictive value of
recurrent stroke in patients with high-grade shunts (Spencer Grade 3 or
more) compared to TEE [17].
Brain magnetic resonance imaging (MRI), including diusion-
weighted imaging (DWI) and magnetic resonance angiography (MRA;
cervical gadolinium-enhanced MRA and three-dimensional time-of-
ight MRA of the circle of Willis) were performed in all patients. FFA
plasma concentration was determined using a colorimetric assay
(ADVIA 1650; Siemens, Erlangen, Germany) using fasting blood ob-
tained either upon the patient's admission, or on the subsequent day.
2.3. Echocardiography parameters
We collected echocardiographic parameters from all patients ac-
cording to the current guidelines [18]. Biplane calculation for left at-
rium (LA) volume was performed using the left atrium areas and
lengths measured from both apical four- and two-chamber views as
guidelines [18]. Calculated LA volumes were indexed to body surface
area (LA volume index, LAVI). Two-dimensional volumetric measure-
ments were performed based on tracings of the blood-tissue interface on
apical four- and two-chamber views at end-diastole and end-systole to
obtain LA areas. End-diastolic areas were digitized from apical biplane
echo images coincident with the beginning of the Q-wave on the elec-
trocardiogram. End-systolic areas were digitized from video frames
near the end of the T-wave with the maximum inward motion of the left
ventricular walls. LA volumes were calculated at end-diastole and end-
systole using Echopac software (GE Vingmed, Milwaukee, Wisconsin)
using the modied Simpson's rule [18].
117
Journal of the Neurological Sciences 382 (2017) 116121
2.4. Acute ischemic lesion analysis
The MRI and MRA parameters used in this study are described in
detail elsewhere [19]. The ischemic brain lesions on DWI were mea-
sured in terms of size, composition, and distribution. The largest dia-
meter of each lesion and the total ischemic lesion volume were mea-
sured using an image-analysis program (Medical Image Processing
Analysis and Visualization; Center for Information Technology, Na-
tional Institutes of Health, Bethesda, MD, USA). We divided the patients
into three infarct composition pattern groups based on the observed
pattern of the ischemic lesion(s): (1) small (< 10 mm in diameter on
DWI) lesions only, (2) mixed small and large lesions, and (3) large le-
sions only [19]. In addition, patients were categorized based on topo-
graphy pattern of the observed ischemic lesion(s): (1) subcortical-only
pattern, (2) small cortical-only pattern, and (3) large cortical/cortical-
deep pattern [20].
2.5. Statistical analysis
Dierences in discrete variables among the groups were examined
using 2, Fisher's exact, or MannWhitney tests. Dierences in con-
tinuous variables were examined using one-way analyses of variance,
KruskalWallis tests, or t-tests. For the statistical analysis, levels of FFA
were grouped into quartiles based on sample size. In addition, in-
dependent factors for ischemic brain lesion volume were evaluated
using linear logistic regression. Adjustment variables in the multi-
variable regression models were chosen from potential outcome de-
terminants with signicant clinical relevance or those that had P-va-
lues < 0.20 in their associations with the outcome in univariate
analysis.
3. Results
Included in the analyses were 254 patients with nonarterial-origin
embolic stroke, 121 with CES, and 133 with ESUS. The patient selection
process is shown in Fig. 1. After extensive embolic source evaluation, it
was determined that the cardiac sources of cerebral emboli in the CES
group were atrial brillation (n = 118, 97.5%), moderate/severe mitral
stenosis (n = 2, 1.7%), and a mechanical prosthetic cardiac valve
(n = 1, 0.8%). In the ESUS group, the presumed possible sources of
embolism were patent foramen ovale in 36 patients (27.1%) and
Fig. 1. Patient selection.
Abbreviations: MRI, magnetic resonance imaging; MRA, magnetic resonance angio-
graphy; CES, cardioembolic stroke; ESUS, embolic stroke of undetermined source.
Extensive embolic source work-ups: repetitive electrocardiography, transthoracic or
transesophageal echocardiography, echo bubble tests or transcranial Doppler right-to-left
shunt tests, and cardiac telemetry (72 h or more).
J.-W. Chung et al. Journal of the Neurological Sciences 382 (2017) 116121

Table 1 Table 2
Baseline characteristics of study participants. Echocardiographic and magnetic resonance imaging characteristics of the CES and ESUS
groups.
CES (n = 121) ESUS (n = 133) P value
CES (n = 121) ESUS (n = 133) P value
Age (years) 71.17 10.51 61.93 15.62 0.001
Male sex, n (%) 31 (41.3) 152 (58.0) 0.011 Echocardiography parameters
Stroke risk factors, n (%) LVIDd (mm) 49.85 5.33 49.73 5.22 0.861
Hypertension 82 (67.8) 65 (48.9) 0.002 LVIDs (mm) 31.61 6.44 30.30 5.26 0.096
Diabetes 34 (28.1) 27 (20.3) 0.146 IVSd (mm) 9.23 1.27 9.54 1.72 0.120
Hypercholesterolemia 44 (36.4) 56 (42.4) 0.325 LVPWd (mm) 8.97 1.09 9.37 1.36 0.017
Atrial brillation 118 (97.5) 0 (0.0) < 0.001 LA (mm) 46.42 7.88 38.94 6.18 < 0.001
Coronary artery disease 58 (47.9) 26 (19.5) < 0.001 Ao (mm) 33.72 4.54 34.03 4.13 0.594
Smoking status 6 (5.0) 12 (9.0) 0.207 LVMI (g/m2) 99.26 24.27 101.25 26.46 0.556
Medications prior to stroke, n RWT 0.36 0.06 0.38 0.07 0.023
(%) DT (msec) 213.59 88.12 240.4 52.79 0.007
Statins 39 (32.2) 38 (28.6) 0.526 LAVI (mL/m2) 64.16 30.24 36.69 12.21 < 0.001
Antiplatelets 45 (37.2) 38 (28.6) 0.114 Diastolic parameters
Warfarin 23 (19.0) 5 (3.8) 0.352 E (m/s) 0.94 0.31 0.65 0.2 < 0.001
NOAC 5 (4.1) 0 (0.0) 0.018 A (m/s) 0.71 0.27 0.76 0.24 0.329
Laboratory ndings E/A 1.34 1.06 0.94 0.45 0.002
Fasting glucose (mg/dL) 127.15 49.21 124.77 39.09 0.668 e (m/s) 0.07 0.02 0.06 0.02 < 0.001
Total cholesterol (mg/dL) 165.61 37.85 176.75 43.40 0.031 a (m/s) 0.07 0.02 0.10 0.09 0.097
Triglycerides (mg/dL) 109.71 65.54 135.10 91.77 0.013 E/e 14.29 7.99 11.59 6.12 0.006
HDL cholesterol (mg/dL) 53.34 15.35 53.13 19.36 0.924 LVEF (%) 0.62 0.06 0.64 0.06 0.013
LDL cholesterol (mg/dL) 100.11 31.50 108.08 39.31 0.078 Size of infarcts
hsCRP (mg/dL) 0.54 1.21 0.66 4.40 0.785 Lesion volume (mm3) 16.37 17.65 7.67 10.55 < 0.001
Free fatty acid (mEq/L) 0.78 0.52 0.67 0.61 0.120 Composition of infarcts, n (%) 0.002
Initial NIHSS, median (IQR) 5 [2 12] 1 [35] < 0.001 Small lesions only 44 (36.4) 75 (56.4)
Small and large lesions, 51 (42.1) 31 (23.3)
Values are mean SD, unless noted otherwise. mixed
Abbreviations: CES, cardioembolic stroke; ESUS, embolic stroke of undetermined source; Large lesions only 26 (21.5) 27 (20.3)
NOAC, novel oral anticoagulants; HDL, high-density lipoprotein; LDL, low-density lipo- Infarct pattern, n (%) < 0.001
protein; hsCRP, high-sensitivity C reactive protein; NIHSS, National Institutes of Health Subcortical-only pattern 4 (3.3) 25 (18.8)
Stroke Scale; IQR, interquartile range; SD, standard deviation. Small cortical-only pattern 40 (33.1) 59 (44.4)
Large cortical/cortical-deep 77 (63.6) 49 (36.8)
patterns
complex aortic arch plaque in 16 (12.0%) patients. After extensive Multiplicitya, median (IQR) 6.0 [1.014.5] 2.0 [1.07.0] < 0.001
etiologic evaluation, no apparent source was identied in 81 (60.9%)
patients with ESUS. Values are mean SD, unless noted otherwise.
The baseline characteristics of the study participants are summar- Abbreviations: CES, cardioembolic stroke; ESUS, embolic stroke of undetermined source;
LVIDs, left ventricular internal diameter end systole; IVSd, interventricular septal end
ized in Table 1. The average age (71.17 10.51 vs. 61.93 15.62,
diastole; LVPWd, left ventricular posterior wall end diastole; LA, left atrium diameter; Ao,
P = 0.001) was higher and the proportion of male patients (41.3% vs. aortic root diameter; LVMI, left ventricular mass index; RWT, relative wall thickness; DT,
58.0%, P = 0.011) was lower in the CES group than in the ESUS group. E wave deceleration time; LAVI, left atrium volume index; E, early wave velocity after
Compared to the ESUS group, the CE group had a higher number of contraction; A, late wave velocity after contraction; e, early velocity of the mitral an-
stroke risk factors and frequency of patients with a prior history of nulus; a, late velocity of the mitral annulus; LVEF, left ventricular ejection fraction; IQR,
hypertension, atrial brillation, and coronary artery disease. Labora- interquartile range; SD, standard deviation.
a
Multiplicity: number of isolated ischemic lesions.
tory test results indicated that the levels of total cholesterol and tri-
glycerides were higher in the ESUS group. However, other biomarker
levels, including FFA (mEq/L, 0.78 0.52 vs. 0.67 0.61, lesions (42.1% vs. 23.3%, P = 0.002) and large cortical/cortical-deep
P = 0.120), were comparable in the two groups. National Institutes of infarct patterns (63.6% vs. 36.8%, P < 0.001) were higher in the CES
Health Stroke Scale (NIHSS) scores (median, 5; interquartile range, group.
212 vs. median, 3; interquartile range, 15; P < 0.001) were higher The quartiles of FFA levels (1Q, 0.030.34; 2Q, 0.350.63; 3Q,
in the CES group, indicating a more severe neurologic presentation. 0.640.91; 4Q, 0.924.51 mEq/L) were strongly associated with the
The details of the echocardiographic ndings for the CES and ESUS characteristics of acute ischemic lesions in nonarterial-origin embolic
groups are summarized in Table 2. Compared to patients with ESUS, infarctions (Table 3). The ischemic lesion volume on DWI increased
those with CES had larger left atrium diameters (46.42 7.88 vs. with FFA quartile (1Q, 5.80 10.11; 2Q, 8.70 10.46; 3Q,
38.94 6.18, P < 0.001) and higher left atrium volume indices 12.80 12.84; 4Q, 19.87 20.44 mm3; P < 0.001). The propor-
(64.16 30.24 vs. 36.69 12.21, P < 0.001). Other parameters, tion of patients with mixed (small and large) ischemic lesions and a
including cardiac morphology, diastolic thickness of the left ventricle large cortical/cortical-deep lesion pattern increased with FFA quartile.
posterior wall, relative wall thickness, and E wave deceleration time Initial stroke severity and echocardiography parameters according to
(DT), were lower in patients with CES. The diastolic function para- quartiles of FFA are summarized in Supplemental Table 1. The initial
meters of E/A (transmitralic ow) ratio (1.34 1.06 vs. 0.94 0.45, NIHSS score increased with quartile of FFA. Among the echocardio-
P = 0.002) and mitral E/e' ratio (14.29 7.99 vs. 11.59 6.12, graphy parameters, only the left atrium volume index showed an in-
P = 0.013) were signicantly higher in the CES group than in the ESUS creasing trend with FFA quartile (P = 0.055).
group. Based on these results, multivariable logistic regression analyses
Acute ischemic lesion(s) detected on DWI also diered among the were performed to determine the independent factors associated with
two groups (Table 2). The volumes of the ischemic lesions ischemic lesion volume in nonarterial-origin embolic stroke patients
(16.37 17.65 vs. 7.67 10.55, P < 0.001) were larger and the (Table 4). Among the clinical factors and laboratory ndings, atrial
number of isolated ischemic lesions (median, 6; interquartile range, brillation (coecient, 7.673; P < 0.001), fasting glucose level
1.014.5 vs. median, 2; interquartile range, 1.07.0; P < 0.001) was (coecient, 0.104; P < 0.001), and FFA level (coecient, 5.249;
higher in the cardioembolism group. In addition, the frequencies of P = 0.001) were positively associated with ischemic lesion volume
patients with ischemic lesions composed of mixed small and large after adjustment for possible confounders. The association between FFA

118
J.-W. Chung et al. Journal of the Neurological Sciences 382 (2017) 116121

Table 3
Ischemic lesion characteristics by quartiles of free fatty acid.

FFA 1Q FFA 2Q FFA 3Q FFA 4Q P value

Size of infarcts, mean SD

Lesion volume (mm3) 5.80 10.11 8.70 10.46 12.80 12.84 19.87 20.44 < 0.001
Composition of infarcts, n (%) 0.002
Small lesions only 40 (63.5) 32 (50.0) 23 (36.5) 24 (37.5)
Small and large lesions, mixed 12 (19.0) 14 (21.9) 26 (41.3) 30 (46.9)
Large lesions only 11 (17.5) 18 (28.1) 14 (22.2) 10 (15.6)
Infarct pattern, n (%) 0.019
Subcortical-only pattern 11 (17.5) 7 (10.9) 7 (11.1) 4 (6.3)
Small cortical-only pattern 31 (49.2) 29 (45.3) 19 (30.2) 20 (31.3)
Large cortical/cortical-deep patterns 21 (33.3) 28 (43.8) 37 (58.7) 40 (62.5)
Multiplicitya, median (IQR) 1.00 [1.005.00] 1.00 [1.006.75] 4.00 [1.0012.00] 9.00 [3.2520.50] < 0.001

Abbreviations: FFA, free fatty acid; Q, quartile; SD, standard deviation; IQR, interquartile range.
a
Multiplicity: number of isolated ischemic lesions.

level and ischemic lesion volume remained signicant when the asso-
ciation was assessed separately in the subgroups of patients with CES
(coecient, 8.884; P = 0.004) and ESUS (coecient, 5.133;
P = 0.001) (Fig. 2). The association between FFA and ischemic lesion
volume was not aected by the cause of the embolism (cardioembolism
vs. ESUS) (P for the interaction between FFA and ischemic lesion vo-
lume, 0.871).
4. Discussion
The proarrhythmic and prothrombotic eect of FFA may explain, at
least in part, the common eect of FFA on ischemic lesion volume in
CES and ESUS in this study. Circulating FFAs, known as nonesteried
fatty acids, are mainly released from adipocyte triglyceride stores of
adipose tissue and serve as physiologically important energy substrates
for the body. Various mechanisms of excess FFA-induced arrhythmia
have been suggested: (1) shortening of action-potential duration by
extracellular K+ accumulation following the modulation of adenosine
triphosphate (ATP)-sensitive K+ channels and ATP-insensitive K+
channels [21], (2) inducing the uncoupling of myocardial cells by
overloading Ca2 + following the perfusion of high molar ratios of al-
bumin-bound FFA [22], (3) activating protein kinases and aecting
Na+, K+, and Ca2 + currents by the inhibition of the Na+/K+ ATPase
pump. In addition to its proarrhythmic eect, FFA increases oxidative
stress [23] and induces hypercoagulability through increased -
brinogen and Factor VII levels [24].
The presence of a potential high-risk cardiac source of embolism in
the absence of signicant arterial disease remains the mainstay for the
clinical diagnosis of CES. However, the risk of cerebral embolism diers
among patients with atrial brillation depending on the patient's
characteristics, such as CHADS2 or CHA2DS2-VASc. Better strategies for
identifying thromboembolic risk in patients with atrial brillation are
needed [25]. In the presence of atrial brillation, other clinical or la-
boratory factors may play a role in the development of embolic stroke.
Elevated FFA levels were found to be associated with recurrent stroke
and poor outcomes in patients with stroke with high-risk cardioembolic
sources or atrial brillation [14,15]. Further studies are needed to
evaluate the roles of the blood biomarkers on infarct characteristics in
patients with embolic stroke of nonarterial origin.
Furthermore, there is increasing evidence that left atrial dysfunction
in the absence of atrial brillation may cause embolic stroke. Systemic
and atrial tissue substrates play a role in thrombogenesis in analogue to
the ventricular cardiopathy observed in myocardial infarction and heart
failure [25]. Left atrial dysfunction may be a marker of incident atrial
brillation, atrial thrombi, and thromboembolic risks of atrial brilla-
tion [26]. In this context, the use of rapidly measurable and reliable
laboratory or echocardiographic biomarkers, such as FFA would be
helpful. In the present study, elevated FFA was associated with

Table 4
Multivariable analysis of factors potentially associated with ischemic lesion volume.

Univariate analysis Multivariable analysis

Coecient SE P value Coecient SE P value

Age (years) 0.154 0.066 0.020 0.003 0.069 0.970

Male sex 3.579 1.885 0.059 3.321 1.747 0.059
Hypertension 2.470 1.903 0.195 0.857 1.934 0.658
Diabetes 3.264 2.197 0.093 3.497 2.382 0.143
Hypercholesterolemia 1.817 1.932 0.348
Coronary artery disease 3.502 1.992 0.080 0.169 1.901 0.929
Atrial brillation 8.613 1.811 < 0.001 7.673 1.855 < 0.001
Smoking status 5.008 3.660 0.172 5.632 3.308 0.090
Prestroke medication
Statins 0.109 2.051 0.958
Antiplatelets 1.552 2.008 0.440
Anticoagulantsa 0.632 2.804 0.822
Fasting glucose (mg/dL) 0.093 0.210 < 0.001 0.104 0.023 < 0.001
Total cholesterol (mg/dL) 0.026 0.023 0.259
Triglycerides (mg/dL) 0.021 0.012 0.077 0.016 0.011 0.134
HDL cholesterol (mg/dL) 0.041 0.054 0.446
LDL cholesterol (mg/dL) 0.020 0.026 0.443
hsCRP (mg/dL) 0.196 0.287 0.495
Free fatty acid (mEq/L) 7.278 1.593 < 0.001 5.249 3.447 0.001

Abbreviations: SE, standard error; HDL, high-density lipoprotein; LDL, low-density lipoprotein; hsCRP, high-sensitivity C reactive protein.
a
Anticoagulants: warfarin or novel oral anticoagulants.

119
J.-W. Chung et al.
echocardiographic ndings of left atrial dysfunction, although the as-
sociation was statistically non-signicant, probably due to the small
number of patients in the cohort.
Numerous studies have investigated diverse strategies to lower
serum FFA levels. A dietary supplement of polyunsaturated fatty acids
was shown to reduce the FFA concentration in plasma and in cell
membranes and exhibited an antiarrhythmic eect [27,28]. In addition,
statins and HMG-CoA reductase inhibitor therapy has been reported to
lower plasma concentrations of FFA, independent of low-density lipo-
protein cholesterol lowering capacity [29]. However, the cardiovas-
cular and metabolic signicance of the FFA lowering eects of statins
requires further investigation. The relation of FFA to insulin resistance,
obesity, and fatty liver disease suggests the important roles of weight
loss and exercise in lowering serum FFA levels, and consequently the
prevention of stroke due to FFA elevation [30]. Further studies in-
vestigating FFA lowering strategies for the prevention and reduction of
stroke severity in embolic stroke are needed.
4.1. Study limitations
Several limitations of this study must be acknowledged. First, the
study population was small and was obtained from a single center.
Therefore, the generalizability of the study results is limited. However,
extensive etiologic evaluations according to institutional protocol for
nonatherogenic embolism sources in a comprehensive stroke care
center may have provided us with a homogenous group of patients and
a lower risk of error in the diagnosis of stroke etiology. Second, the
overnight fasting plasma level of FFA was measured at a single time
point in the present study, at a median of 2 days after stroke onset
(range, 04 days). Single-point fasting FFA may not adequately reect
circadian variations in FFA. FFA concentrations strongly depend on the
nutritional state, with generally lower FFA levels observed the post-
prandial period due to the insulin-mediated suppression of lipolysis
[31]. Third, even though nonarterial-origin embolic stroke patients
were enrolled prospectively in the present study, the plasma FFA levels
were measured only at a single time point. Therefore, a direct causal
relationship linking FFA with ischemic lesions pattern and volume
could not be determined in this study. A future study with serial plasma
FFA measurements is warranted. Lastly, total FFA comprises a variety of
fatty acids of diering hydrocarbon lengths and saturation levels. Data
regarding specic FFAs, including trans fatty acids, which are known to
adversely aect cardiovascular risk, were not available for this study.
Future studies examining the impacts of individual FFAs on acute is-
chemic lesion characteristics are warranted.
120
Journal of the Neurological Sciences 382 (2017) 116121
Fig. 2. Ischemic lesion volume by FFA levels for the CES and
ESUS groups.
Abbreviations: FFA, free fatty acid; CES, cardioembolic stroke;
ESUS, embolic stroke of undetermined source; CI, condence
interval.
5. Conclusions
In conclusion, our data indicate that high plasma FFA levels are
associated with increased ischemic lesion volume in both CES and
ESUS. In addition, increased FFA levels correlated with more severe
initial neurologic decits in nonatherogenic embolic infarctions. Our
results emphasize the importance of identifying common biomarkers
for the identication of thromboembolic risk and developing FFA
lowering strategies to reduce the degree of ischemic insult in patients
with CES and ESUS.
Source of funding
This study was supported by the Korean Healthcare Technology
R & D Project, Ministry of Health & Welfare (HI14C1531).
Disclosures
All authors report no potential conicts of interest.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jns.2017.09.040.
References
[1] K.L. Furie, S.E. Kasner, R.J. Adams, G.W. Albers, R.L. Bush, S.C. Fagan, et al.,
Guidelines for the prevention of stroke in patients with stroke or transient ischemic
attack: a guideline for healthcare professionals from the American Heart
Association/American Stroke Association, Stroke 42 (2011) 227276.
[2] R.L. Sacco, J.H. Ellenberg, J.P. Mohr, T.K. Tatemichi, D.B. Hier, T.R. Price, et al.,
Infarcts of undetermined cause: the NINCDS Stroke Data Bank, Ann. Neurol. 25
(1989) 382390.
[3] P. Amarenco, Cryptogenic stroke, aortic arch atheroma, patent foramen ovale, and
the risk of stroke, Cerebrovasc. Dis. 20 (2005) 6874.
[4] O.Y. Bang, B. Ovbiagele, J.S. Kim, Evaluation of cryptogenic stroke with advanced
diagnostic techniques, Stroke 45 (2014) 11861194.
[5] H. Ay, K.L. Furie, A. Singhal, W.S. Smith, A.G. Sorensen, W.J. Koroshetz, An evi-
dence-based causative classication system for acute ischemic stroke, Ann. Neurol.
58 (2005) 688697.
[6] D.M. Kent, R. Ruthazer, C. Weimar, J.L. Mas, J. Serena, S. Homma, et al., An index
to identify stroke-related vs incidental patent foramen ovale in cryptogenic stroke,
Neurology 81 (2013) 619625.
[7] A.O. Ozdemir, A. Tamayo, C. Munoz, B. Dias, J.D. Spence, Cryptogenic stroke and
patent foramen ovale: clinical clues to paradoxical embolism, J. Neurol. Sci. 275
(2008) 121127.
[8] M.R. Di Tullio, C. Russo, Z. Jin, R.L. Sacco, J.P. Mohr, S. Homma, et al., Aortic arch
plaques and risk of recurrent stroke and death, Circulation 119 (2009) 23762382.
[9] W.E. Connor, The acceleration of thrombus formation by certain fatty acids, J. Clin.
Invest. 41 (1962) 11991205.
[10] J.C. Hoak, E.D. Warner, W.E. Connor, Platelets, fatty acids and thrombosis, Circ.
Res. 20 (1967) 1117.
J.-W. Chung et al.
[11] H. Bays, L. Mandarino, R.A. DeFronzo, Role of the adipocyte, free fatty acids, and
ectopic fat in pathogenesis of type 2 diabetes mellitus: peroxisomal proliferator-
activated receptor agonists provide a rational therapeutic approach, J. Clin.
Endocrinol. Metab. 89 (2004) 463478.
[12] L.H. Opie, The metabolic vicious cycle in heart failure, Lancet 364 (2004)
17331734.
[13] M.F. Oliver, L.H. Opie, Eects of glucose and fatty acids on myocardial ischaemia
and arrhythmias, Lancet 343 (1994) 155158.
[14] J.Y. Choi, J.S. Kim, J.H. Kim, K. Oh, S.B. Koh, W.K. Seo, High free fatty acid level is
associated with recurrent stroke in cardioembolic stroke patients, Neurology 82
(2014) 11421148.
[15] J.Y. Choi, J.M. Jung, D.Y. Kwon, M.H. Park, J.H. Kim, K. Oh, et al., Free fatty acid as
an outcome predictor of atrial brillation-associated stroke, Ann. Neurol. 79 (2016)
317325.
[16] R.G. Hart, H.C. Diener, S.B. Coutts, J.D. Easton, C.B. Granger, M.J. O'Donnell, et al.,
Embolic strokes of undetermined source: the case for a new clinical construct,
Lancet Neurol. 13 (2014) 429438.
[17] J. Tobe, C. Bogiatzi, C. Munoz, A. Tamayo, J.D. Spence, Transcranial doppler is
complementary to echocardiography for detection and risk stratication of patent
foramen ovale, Can. J. Cardiol. 32 (2016) (986 e989986 e916).
[18] R.M. Lang, L.P. Badano, V. Mor-Avi, J. Alalo, A. Armstrong, L. Ernande, et al.,
Recommendations for cardiac chamber quantication by echocardiography in
adults: an update from the American Society of Echocardiography and the European
Association of Cardiovascular Imaging, J. Am. Soc. Echocardiogr. 28 (2015) 139
(e14).
[19] O.Y. Bang, J.W. Chung, S. Ryoo, G.J. Moon, G.M. Kim, C.S. Chung, et al., Brain
microangiopathy and macroangiopathy share common risk factors and biomarkers,
Atherosclerosis 246 (2016) 7177.
121
Journal of the Neurological Sciences 382 (2017) 116121
[20] O.Y. Bang, B. Ovbiagele, D.S. Liebeskind, L. Restrepo, S.R. Yoon, J.L. Saver, Clinical
determinants of infarct pattern subtypes in large vessel atherosclerotic stroke, J.
Neurol. 256 (2009) 591599.
[21] D. Kim, R.A. Du, Regulation of K+ channels in cardiac myocytes by free fatty
acids, Circ. Res. 67 (1990) 10401046.
[22] A.F. Willerbrands, H.F. ter Welle, S.J. Tasseron, The eect of a high molar FFA-
albumin ratio in the perfusion medium on rhythm and contractility of the isolated
rat heart, J. Mol. Cell. Cardiol. 5 (1973) 259273.
[23] G. Soardo, D. Donnini, L. Domenis, C. Catena, D. De Silvestri, D. Cappello, et al.,
Oxidative stress is activated by free fatty acids in cultured human hepatocytes,
Metab. Syndr. Relat. Disord. 9 (2011) 397401.
[24] M. Mutanen, R. Freese, Fats, lipids and blood coagulation, Curr. Opin. Lipidol. 12
(2001) 2529.
[25] H. Kamel, P.M. Okin, M.S. Elkind, C. Iadecola, Atrial brillation and mechanisms of
stroke: time for a new model, Stroke 47 (2016) 895900.
[26] D.Y. Leung, A. Boyd, A.A. Ng, C. Chi, L. Thomas, Echocardiographic evaluation of
left atrial size and function: current understanding, pathophysiologic correlates,
and prognostic implications, Am. Heart J. 156 (2008) 10561064.
[27] A. Leaf, P.C. Weber, Cardiovascular eects of n-3 fatty acids, N. Engl. J. Med. 318
(1988) 549557.
[28] K.H. Weylandt, J.X. Kang, A. Leaf, Polyunsaturated fatty acids exert antiarrhythmic
actions as free acids rather than in phospholipids, Lipids 31 (1996) 977982.
[29] A. Sahebkar, L.E. Simental-Mendia, C. Pedone, G. Ferretti, P. Nachtigal, S. Bo, et al.,
Statin therapy and plasma free fatty acids: a systematic review and meta-analysis of
controlled clinical trials, Br. J. Clin. Pharmacol. 81 (2016) 807818.
[30] J.D. Spence, Secondary stroke prevention, Nat. Rev. Neurol. 6 (2010) 477486.
[31] K.N. Frayn, Plasma non-esteried fatty acids: why are we not measuring them
routinely? Ann. Clin. Biochem. 42 (2005) 413414.